Year: 2025
Type: article
Abstract: Generative AI, particularly large language models (LLMs), is reshaping the cybersecurity landscape by enabling both innovative defense mechanisms and novel forms of attack. This article explores the dual role of generative AI in both offensive and defensive cybersecurity operations. While GenAI offers significant advancements in defensive capabilities, it is also being leveraged by nation-state actors to enhance the sophistication and success rates of cyberattacks. The article analyzes how LLMs are applied in offensive engagements such as red teaming, penetration testing, and threat intelligence, while also identifying emerging technical, operational, and strategic risks associated with their deployment. Special attention is given to the cybersecurity challenges of generative AI systems themselves, highlighting limitations in conventional frameworks and proposing governance-oriented mitigations such as model evaluation, human-in-the-loop oversight, GenAI-specific red teaming, and the structured dissemination of threat intelligence derived from GenAI-enabled security practices.
Author: Kunter Orpak
Institution: University of Amsterdam
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
Funder: Universiteit van Amsterdam
OpenAlex ID: W4414092634